Enhanced Tribological, Corrosion, and Microstructural Properties of an

May 29, 2014 - alloy-based magnetic recording media which is used to store digital information. To protect the media from mechanical wear and oxidatio...
0 downloads 0 Views 5MB Size
Research Article www.acsami.org

Enhanced Tribological, Corrosion, and Microstructural Properties of an Ultrathin (500 °C) and rapid thermal cycling rates. Hence, the thermal stability of the protective overcoat is also one of the important issues that need to be addressed. Filtered cathodic vacuum arc (FCVA) is one of the techniques which can deposit carbon overcoats (COCs) on media with enhanced protective characteristics, low surface roughness, and good film coverage even at lower thicknesses.7−10 FCVA generates a highly ionized plasma (∼90%) of energetic C+ ions, which helps to achieve a high amount of sp3 hybridization of carbon within the COC. The high sp3 carbon (sp3C) fraction directly contributes to its hardness, elastic modulus, density, and wear resistance, and indirectly contributes to its coverage excellence.9,11 In addition, FCVAdeposited COC on magnetic media has demonstrated higher thermal stability as compared to a conventional PECVD deposited hydrogenated COC (the thicknesses of the COCs

1. INTRODUCTION Commercial hard disk drives (HDDs) today contain a Coalloy-based magnetic recording media which is used to store digital information. To protect the media from mechanical wear and oxidation, the media is coated with a thin, hard, and chemically inert layer of diamond-like carbon (DLC), which is topped with a monolayer of lubricant to achieve low friction at the head−disk interface. To achieve higher areal densities, the spacing between the magnetic media and the read head, known as the magnetic spacing, has to be further reduced.1−3 This has motivated a continuous decrease in the DLC overcoat thickness. Future generation HDDs with areal densities > 1 Tb/in2 require a protective overcoat thickness of 3.5 nm).12,13 The higher thermal stability of the FCVA-deposited COC, which is non-hydrogenated, has been attributed to the higher sp3C fraction which can be obtained by the FCVA deposition technique. As such, it can be seen that the advantage of the FCVA deposition technique lies in its ability to attain a high sp3C fraction, which provides all the desirable yet essential qualities for future hard disk media overcoats. However, at ultrathin levels, FCVA-deposited carbon films show significantly lower sp3C content as compared to its thicker counterparts. For example, while FCVA-deposited carbon can have up to ∼85% sp3C content for 200 nm films,14 Ferrari has suggested that FCVA-deposited carbon films achieve only ∼50% sp3C content at ∼2 nm thickness.15 Hence, for overcoat thickness < 2 nm, maintaining or improving the sp3C fraction and its associated properties (in terms of its good wear resistance, corrosion resistance, and thermal stability) even with FCVA-deposited carbon would be very challenging. Apart from high sp3C content, high interfacial strength/ bonding at the magnetic media−overcoat interface is also needed for enhanced tribological properties and wear protection. By controlling the C+ ion energy using the FCVA technique, surface modification of the magnetic media had been performed to obtain an ultrathin graded protective overcoat (≤2.0 nm) by forming a mixed interfacial layer of carbon and media.16,17 This graded overcoat structure was seen to improve the interfacial bonding between the COC and media, and exhibited a low coefficient of friction even without a lubricant layer. However, the overcoat fabricated by this FCVA surface modification technique has not been found to improve the corrosion performance over commercial hard disks.18

The addition of an underlayer is another method of improving the interfacial bonding and adhesion. Rismani et al. found that the addition of a Si underlayer for FCVAdeposited carbon overcoats on Co magnetic media formed interfacial Co−Si and Si−C bonds, giving a better wear durability of the overcoat.19 Amorphous silicon nitride (SiNx) films are commonly known for their high hardness, high electrical resistivity, chemical inertness, and good diffusion barrier properties owing to their dense, covalently bonded structure.20 Furthermore, when used as an underlayer, they have been reported to provide good adhesion between DLC and metallic substrates.21 However, SiNx itself is not as wear resistant as DLC, and it is prone to oxidation in ambient oxygen and humidity, especially at low thicknesses.22,23 Recently, Bunnak et al. prepared 10 nm thick composite SiNx/DLC films on Si substrates by combining RF sputtering and FCVA deposition techniques, and have observed promising properties in terms of high sp3 carbon bonding.24 By combining the ideal properties of SiNx and FCVAdeposited carbon, the aim of this study is as follows: (1) to maintain or improve the sp3C bonding within the COC; (2) to provide good overcoat adhesion to the underlying magnetic media; and most importantly (3) to enhance the tribological properties and corrosion resistance of the overcoat. Here in this work, a SiNx/C bilayer overcoat is developed with a total thickness of ∼16 Å showing remarkable properties. To demonstrate the superiority of the SiNx/C bilayer overcoat, a monolithic SiNx overcoat and a monolithic FCVA-deposited carbon overcoat (both with thickness of ∼16 Å) were also prepared to compare their properties. 9377

dx.doi.org/10.1021/am501760p | ACS Appl. Mater. Interfaces 2014, 6, 9376−9385

ACS Applied Materials & Interfaces

Research Article

The schematic of the resultant disk sample is shown in Figure 1c. For the purpose of evaluating the performance of the SiNx/C bilayer overcoat, single layer overcoats of reactive sputtered SiNx and FCVAdeposited carbon, each with thickness of 16 Å, were also deposited on similar etched magnetic media substrates. The performance of each of these samples was compared to a reference commercial media disk sample with its commercial COC but without the lubricant layer. For completeness, a specially prepared commercial disk without any protective overcoat over the magnetic media was also used in this study to provide a benchmark to show the effectiveness of using a protective overcoat on commercial magnetic media. A list of the samples and their nomenclature (which will be used henceforth) is presented in Table 1.

2. MATERIALS AND METHODS 2.1. Sample Preparation. Figure 1 describes the sample preparation process for fabricating the SiNx/C bilayer overcoat on commercial media disks. Commercial 2.5″ disks were used as the starting substrates in this work, consisting of (from bottom to top) a glass disk substrate, a multilayer structure of various materials for optimized recording performance, a CoCrPt:Oxide magnetic recording layer, a COC layer, and finally a lubricant layer. A cross section schematic of the commercial disk structure is provided in Figure 1a. Prior to deposition of the SiNx/C overcoat, the pre-existing commercial COC and lubricant layer were removed by Ar+ ion beam etching at an ion energy of 300 eV, as seen in Figure 1b. A secondary ion mass spectrometer (SIMS) detector was used to calibrate the etching rate to remove the commercial COC. It should be noted that this etching process was applied as a necessary step when fabricating these overcoats for our experiments. However, it is not required in a conventional hard disk manufacturing process. Deposition of the SiNx and carbon films on the etched commercial disk substrates were carried out in situ after etching, using a VEECO deposition system equipped with a pulsed filtered cathodic arc source, sputtering source, and Ar+ ion beam etching capability. A 99.99% pure silicon target was used for the deposition of SiNx, while a 99.999% pure graphite rod was used for the deposition of carbon by FCVA. Deposition was carried out at a background pressure of ∼10−7 Torr. First, a 4 Å ultrathin layer of SiNx was deposited by pulsed DC reactive sputtering in a gaseous mixture of Ar + N2, with a ratio of 67% Ar to 33% N2 and at a duty cycle of 0.7. Next, the sample was transferred to the FCVA chamber in situ under vacuum. A schematic of a FCVA setup for deposition of COC is shown in Figure 2. A 12 Å layer of

Table 1. Description and Nomenclature of Samples Used in This Work nomenclature 16C 16SiN 4SiN12C CM BM

sample type Ar+ etched commercial disk (COC removed) Ar+ etched commercial disk (COC removed) Ar+ etched commercial disk (COC removed) as-received commercial disk (without lube) as-received commercial disk (without COC and without lube)

overcoat structure C (16 Å) SiNx (16 Å) SiNx (4 Å)|C (12 Å) commercial COC (∼27 Å) no overcoat (bare media)

2.2. Characterization Methods. High resolution cross-sectional transmission electron microscopy (TEM, Philips FEG CM300) was performed to image the microstructure and thickness of the overcoats after deposition. Before imaging, all samples (16C, 16SiN, 4SiN12C, and CM) were prepared for cross-sectional TEM measurements. To investigate any surface modification induced changes in the smoothness of the media surface, the surface roughnesses of samples S-1 to S-4 were measured with the help of tapping mode atomic force microscopy (AFM, Bruker Innova). The measurements were conducted at a scan area of 2 μm × 2 μm at three different points on each sample surface, from which a mean value was taken. One of the methods to explore the wear resistance of the overcoats on magnetic media is to investigate their tribological properties through ball-on-disk tribological tests. Ball-on-disk tribological tests were carried out on all the samples using a nanotribometer (CSM Instruments). Sapphire (Al2O3) was used as the counterface ball material, and a contact load of 20 mN (the minimum load which can be applied by the tribometer) was kept constant in all the tests. The sample was rotated such that the ball slid across the sample surface in a circular motion with a radius of 1.2 mm and at a linear speed of 1.0 cm s−1. Each test was carried out on at least two locations on each sample for up to 10 000 cycles while the coefficient of friction, μ, was measured. After the test, the wear track and ball images were captured by using an optical microscope. During the surface modification process, the C+ ions may also interact with the magnetic recording media. Hence, it is crucial to investigate any change in the macromagnetic properties of the magnetic media. The macromagnetic properties of the disk samples after surface modification were measured using a custom-made magneto-optic Kerr effect (MOKE) setup to investigate whether the surface modification process had significantly affected the magnetic performance of the magnetic media. The hysteresis loops of Kerr rotation with respect to the applied magnetic field for samples 16C, 4SiN12C, and CM were recorded and compared. At ultralow thicknesses, protection of the underlying hard disk media from corrosion or oxidation becomes critical to prevent the degradation of the magnetic material over time which would lead to the loss of stored data. To understand the effectiveness of the overcoats in protecting the underlying media from corrosion, a custom-made three-electrode corrosion setup was used to perform

Figure 2. Schematic of a typical FCVA setup for the deposition of COC. carbon was subsequently deposited above the SiNx layer by pulsed FCVA deposition, giving a total overcoat thickness of 16 Å. During the FCVA deposition process, an arc was struck between the anode and cathode using a high current pulsed power supply with a duty cycle of 0.001. The resulting arc discharge was transported using a single 90° bend filter coil which magnetically confines and guides the plasma toward the substrate. This filter helps to remove any neutrals or macroparticles in the plasma which do not react with the coil’s magnetic field. The plasma exiting the filter coil then passes through a plasma shaping coil before reaching the substrate. No substrate bias was applied during the FCVA deposition process, and hence, the average energy of the C+ ions arriving at the substrate was around 25 eV.25 The deposition rates of both the sputtered SiNx and FCVAdeposited carbon layers were calibrated individually by X-ray reflectivity (XRR). The deposition rate of carbon by FCVA was found to be 0.063 Å/pulse. By adjusting the number of pulses, the COCs of desired thicknesses were deposited in samples 16C and 4SiN12C. In addition, the etching rate uniformity as well as the thickness uniformity of the SiNx and C films were qualified prior to deposition. 9378

dx.doi.org/10.1021/am501760p | ACS Appl. Mater. Interfaces 2014, 6, 9376−9385

ACS Applied Materials & Interfaces

Research Article

Figure 3. Cross-sectional TEM images showing the thickness of the overcoats (labeled in red) for samples (a) 16C, (b) 16SiN, (c) 4SiN12C, and (d) CM.

4SiN12C were each measured to be ∼1.6 ± 0.1 nm, while the thickness of the COC in sample CM was measured to be 2.7 ± 0.1 nm. It is evident that the deposited overcoat thicknesses in samples 16C, 16SiN, and 4SiN12C matched well with our calibration, revealing that our process can be precisely controlled even at ultrathin levels. In sample 4SiN12C, however, owing to the extremely low thickness of the SiNx layer and similar contrast to carbon, it is difficult to distinguish it from the FCVA-deposited carbon layer. Consequently, only the total overcoat thickness of ∼1.6 ± 0.1 nm was measured. The presence of SiNx in sample 4SiN12C was subsequently confirmed by XPS analysis, which will be discussed later. The average roughness (Ra) and root-mean-square roughness (Rq) of the four samples were measured, as shown in Figure 4. It can be seen that the Rq of the COC in commercial disks

electrochemical potentiodynamic polarization measurements on all the samples.18 An electrolyte solution of 0.1 M NaCl was used during the test, and a geometric surface area of 0.24 cm2 of the sample surface was exposed to the electrolyte. Each electrochemical test consisted of an anodic sweep and a cathodic sweep where the potential was varied and the corresponding current was measured. In the anodic sweep, the potential was swept 0.4 V above the open circuit potential, whereas in the cathodic sweep the potential was swept 0.4 V below the open circuit potential. Every sweep was conducted at a different location on the sample, with at least three sets of tests (6 sweeps) conducted on each sample to obtain consistent readings. The test best representing the consistent result is presented. X-ray photoelectron spectroscopy (XPS) was used to characterize the chemical bonding and oxidation resistance performance of the carbon-containing overcoats, namely, 16C, 4SiN12C, and CM. XPS measurements were performed using a VG ESCALAB 220I-XL spectrometer with an Al Kα source. The microstructures of the carbon overcoats in these samples were probed with visible and UV Raman spectroscopy (Jobin Yvon LABRAM-HR) at laser excitation wavelengths of 488 and 325 nm, respectively. To avoid damage to ultrathin COCs due to the laser heating of the sample surface, the laser power was kept low and other parameters such as the charge coupled device’s (CCD) exposure and data acquisition time were synchronized to obtain a reasonable signal-to-noise ratio for all samples. The relative intensity ratios of the COCs were compared from various locations of the sample surfaces to distinguish the nature of the carbon bonding with an accuracy of